Thread (6 messages) 6 messages, 2 authors, 2017-08-04

[PATCH 1/2] perf/aux: Make aux_{head, wakeup} ring_buffer members long

From: Will Deacon <hidden>
Date: 2017-08-04 14:58:36
Also in: lkml

On Mon, Jul 31, 2017 at 01:28:01PM +0300, Alexander Shishkin wrote:
Will Deacon [off-list ref] writes:
quoted
The aux_head and aux_wakeup members of struct ring_buffer are defined
using the local_t type, despite the fact that they are only accessed via
the perf_aux_output_* functions, which cannot race with each other for a
given ring buffer.

This patch changes the type of the members to long, so we can avoid
using the local_* API where it isn't needed.
Thanks for digging this up! Some minor nits below.
quoted
@@ -434,12 +434,12 @@ void perf_aux_output_end(struct perf_output_handle *handle, unsigned long size)
 		handle->aux_flags |= PERF_AUX_FLAG_OVERWRITE;
 
 		aux_head = handle->head;
-		local_set(&rb->aux_head, aux_head);
+		rb->aux_head = aux_head;
 	} else {
 		handle->aux_flags &= ~PERF_AUX_FLAG_OVERWRITE;
 
-		aux_head = local_read(&rb->aux_head);
-		local_add(size, &rb->aux_head);
+		aux_head = rb->aux_head;
+		rb->aux_head += size;
 	}
 
 	if (size || handle->aux_flags) {
@@ -451,11 +451,11 @@ void perf_aux_output_end(struct perf_output_handle *handle, unsigned long size)
 		                     handle->aux_flags);
 	}
 
-	aux_head = rb->user_page->aux_head = local_read(&rb->aux_head);
+	aux_head = rb->user_page->aux_head = rb->aux_head;
 
-	if (aux_head - local_read(&rb->aux_wakeup) >= rb->aux_watermark) {
+	if (aux_head - rb->aux_wakeup >= rb->aux_watermark) {
Can't we just do away with aux_head here:

  rb->user_page->aux_head = rb->aux_head;
  if (rb->aux_head - rb->aux_wakeup >= rb->aux_watermark) { ...

?
quoted
@@ -485,14 +485,13 @@ int perf_aux_output_skip(struct perf_output_handle *handle, unsigned long size)
 	if (size > handle->size)
 		return -ENOSPC;
 
-	local_add(size, &rb->aux_head);
+	rb->aux_head += size;
 
-	aux_head = rb->user_page->aux_head = local_read(&rb->aux_head);
-	if (aux_head - local_read(&rb->aux_wakeup) >= rb->aux_watermark) {
+	aux_head = rb->user_page->aux_head = rb->aux_head;
+	if (aux_head - rb->aux_wakeup >= rb->aux_watermark) {
 		perf_output_wakeup(handle);
-		local_add(rb->aux_watermark, &rb->aux_wakeup);
-		handle->wakeup = local_read(&rb->aux_wakeup) +
-				 rb->aux_watermark;
+		rb->aux_wakeup += rb->aux_watermark;
+		handle->wakeup = rb->aux_wakeup + rb->aux_watermark;
 	}
 
 	handle->head = aux_head;
And here I think we don't need aux_head at all.
Agreed on both counts. Will fix for v2.

Will
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help